Journ

alof

Cell

Scie

nce

The genetic depletion or the triptolide inhibition ofTFIIH in p53-deficient cells induces a JNK-dependentcell death in Drosophila

Claudia Villicana, Grisel Cruz and Mario Zurita*Departamento de Genetica del Desarrollo, Instituto de Biotecnologıa, Avenida Universidad 2001, Cuernavaca Morelos, 62250, Mexico

*Author for correspondence (marioz@ibt.unam.mx)

Accepted 7 March 2013Journal of Cell Science 126, 2502–2515� 2013. Published by The Company of Biologists Ltddoi: 10.1242/jcs.122721

SummaryTranscription factor IIH (TFIIH) participates in transcription, nucleotide excision repair and the control of the cell cycle. In the presentstudy, we demonstrate that the Dmp52 subunit of TFIIH in Drosophila physically interacts with the fly p53 homologue, Dp53. The

depletion of Dmp52 in the wing disc generates chromosome fragility, increases apoptosis and produces wings with a reduced number ofcells; cellular proliferation, however, is not affected. Interestingly, instead of suppressing the apoptotic phenotype, the depletion of Dp53in Dmp52-depleted wing disc cells increases apoptosis and the number of cells that suffer from chromosome fragility. The apoptosisinduced by the depletion of Dmp52 alone is partially dependent on the JNK pathway. In contrast, the enhanced apoptosis caused by the

simultaneous depletion of Dp53 and Dmp52 is absolutely JNK-dependent. In this study, we also show that the anti-proliferative drugtriptolide, which inhibits the ATPase activity of the XPB subunit of TFIIH, phenocopies the JNK-dependent massive apoptoticphenotype of Dp53-depleted wing disc cells; this observation suggests that the mechanism by which triptolide induces apoptosis in p53-

deficient cancer cells involves the activation of the JNK death pathway.

Key words: Apoptosis, Drosophila, JNK, TFIIH, Triptolide, p53

IntroductionTo carry out their function, several multisubunit complexes

involved in transcription and/or DNA repair establish multiple

interactions with other factors. This interplay between different

components is highly dynamic, and in many cases, the

interactions are transitory although fundamental for different

cellular processes. A typical example is the 10-subunit complex

TFIIH, which participates in RNA pol II- and RNA pol I-

mediated transcription, in nucleotide excision repair (NER) and

in the control of the cell cycle (Zurita and Merino, 2003). It has

been reported that during transcription by RNA pol II, TFIIH

must interact with different components of the pre-initiation

complex, including TFIIE, subunits of the mediator and RNA pol

II (Egly and Coin, 2011). It has also been reported that TFIIH

interacts with different transcriptional activators (Esnault et al.,

2008; Chymkowitch et al., 2011). These interactions are essential

for the primary functions of TFIIH in transcription, including the

formation of the DNA bubble at the transcription initiation site

that is catalysed by the 39-to-59 helicase activity of the XPB

subunit and the ATPase activity of XPD. In addition, the

phosphorylation of serines 5 and 7 of the CTD domain of the

large subunit of RNA pol II, which is necessary to initiate and

elongate transcription, is catalysed by the transcription-specific

kinase module (CAK) of TFIIH, composed of Cdk7, CycH and

MAT1 (Akhtar et al., 2009; Glover-Cutter et al., 2009). In NER,

TFIIH is recruited to the damaged site by the DDB and XPC-

HR23-b centrin complex in untranscribed regions and by the

stalled RNA pol II in transcribed chromatin (Coin et al., 2008).

The ATPase activity of XPB is required to anchor TFIIH to the

damaged DNA, and the XPD helicase melts the DNA in the 59 to

39 direction, thus allowing the incision of damaged strand by

XPG and ERCC1-XPF endonucleases. TFIIH maintains a close

interaction with XPG and XPA, which stabilise the NER complex

(Egly and Coin, 2011). In fact, it has been suggested that XPG

forms a complex with TFIIH that participates not only in NER

but also in transcription (Ito et al., 2007).

A particularly relevant issue related to TFIIH is its association

with human diseases: mutations in its XPB, XPD and p8 subunits

are linked to xeroderma pigmentosum, Cockayne syndrome,

trichothiodystrophy and cancer (Bergoglio and Magnaldo, 2006).

Mutations in other subunits of TFIIH associated with syndromes

have not yet been reported in humans. Using Drosophila

melanogaster as a model to reveal the function of other

subunits, it has been demonstrated that Dmp52 mutations

generate phenotypes associated with these three syndromes,

indicating that Dmp52 plays an important role in regulating the

entire activity of TFIIH during several cellular processes (Coin

et al., 2007; Fregoso et al., 2007).

One of the transcriptional activators that have been shown to

interact with TFIIH is the tumour suppressor p53 (Leveillard

et al., 1996). After genotoxic stress, p53 is activated and can

induce either cell cycle arrest, DNA repair or direct the cell

toward apoptosis, depending on the severity of the DNA damage

(Brady and Attardi, 2010). The transactivation domain (TAD) of

p53 can interact with the PH domain of the p62 subunit of TFIIH

in human cells; this interaction is correlated with the ability of

2502 Research Article

Journ

alof

Cell

Scie

nce

these proteins to activate both the initiation and elongation oftranscription (Okuda et al., 2008; Di Lello et al., 2008). In

addition, the CAK complex phosphorylates p53 and enhances itsDNA-binding activity (Ko et al., 1997). For UV-induced DNAdamage, p53 is required for the recruitment of TFIIH to the

damaged DNA sites (Wang et al., 1996). Genetically, it has beendemonstrated that in human cells derived from patients whoseXPD and XPB subunits have been mutated, the apoptotic effectcaused by p53 overexpression is suppressed (Wang et al., 1996).

In Drosophila, a similar result was obtained in XPB (haywire)mutant flies (Fuller et al., 1989; Merino et al., 2002). Althoughthese results suggest that p53 requires the presence of intact

TFIIH to promote apoptosis, these experiments were conductedin an artificial situation in which p53 was overexpressed. Inaddition to the established interactions between p62 and the XPB

and XPD subunits of TFIIH, a previous report suggested that p53may also interact with a 52-kDa protein that is present in theTFIIH complex (Leveillard et al., 1996); while this protein wasnot characterised at the time of the initial publication of its

existence, the obvious candidate is the p52 subunit of TFIIH.

The reports of interactions between TFIIH and p53 suggest that

these interactions play an important role at some point during thecell cycle and during the processes that occur after DNA damage.However, there is still limited information on the cross talkbetween p53 and TFIIH during animal development. Here, we

present evidence that the fly homologue of p53 (Dp53) directlyinteracts with the fly p52 (Dmp52) subunit of TFIIH in theabsence of genotoxic stress. The depletion of Dmp52 in the wing

disc caused growth and differentiation defects duringdevelopment as well as chromosomal aberrations that caninduce apoptosis. Intriguingly, the simultaneous depletion of

TFIIH and Dp53 increased the presence of chromosomalaberrations that activate apoptosis in a JNK-dependent manner.Furthermore, the specific inhibition of XPB by the natural

product triptolide phenocopies the apoptotic response caused bythe absence of p52. This discovery has important implications forthe treatment of cell tumours deficient in p53.

ResultsDp53 physically interacts with the Dmp52 subunit of TFIIH

It has been reported that p53 interacts with the p62 subunit ofTFIIH through its PH domain (Okuda et al., 2008; Di Lello et al.,2008). Intriguingly, published data also point to the existence of a

possible direct contact between p53 and the p52 subunit of TFIIH(Leveillard et al., 1996). To gain insight into the physicalinteractions between Dp53 and Dmp52 in Drosophila, we

constructed recombinant Dp53 short (DDNp53) and large(Dp53) isoform proteins (Bourdon et al., 2005) with FLAG orV5 tags at the C-terminal end and Dmp52 tagged at the N-

terminal end (FLAG-Dmp52). These constructions wereexpressed in S2R+ cells, and the expression of the recombinantDp53 and Dmp52 proteins were verified by western blots

and immunoprecipitation assays using specific antibodies(supplementary material Fig. S1).

Using total protein extracts from cells transfected with the

Dp53-tag constructs, we performed co-immunoprecipitation(CoIP) assays. CoIP experiments using whole cell extractsagainst Dp53-FLAG and the DDNp53-V5 recombinant proteins

enabled the pulldown of the endogenous cellular Dmp52(Fig. 1A,B, upper panels). In fact, by using a specific Dp53antibody that recognises both isoforms (supplementary material

Fig. S5), CoIP of the endogenous Dmp52 using S2R+ total

protein cell extracts was possible (Fig. 1A, lower panel). In the

case of the DDNp53-V5 immunoprecipitation experiments, we

also analysed whether Cdk7 and DmXPB (other components

of the TFIIH complex) immunoprecipitated with Dp53-V5.

Surprisingly, we detected no Cdk7 and only very low levels of

DmXPB (Fig. 1B). However, we cannot dismiss the idea that the

location of the V5-epitope on the target protein may affect the

CoIP results or our experimental conditions, thereby impeding or

disaggregating the entire TFIIH complex.

Reciprocal CoIPs using DmXPB, Dp53 or Cdk7 antibodies on

non-transfected S2R+ cells showed that the XPB antibody was

able to co-immunoprecipitate Dp53 and p62 but not Cdk7

(Fig. 1C; supplementary material Fig. S1C). In addition, we did

not detect any CoIP of Cdk7 with either the Dp53 or XPB

antibodies. We have observed similar results in the protein

extracts of these cells using antibodies against other TFIIH core

subunits (unpublished results). Furthermore, CoIPs conducted

with Cdk7 antibodies were not able to immunoprecipitate other

analysed components of TFIIH (Fig. 1C). It is possible that the

conditions used in our CoIP experiments were very strong, which

could have resulted in a disrupted interaction between the CAK

complex and the core of TFIIH. At this point, we decided to

Fig. 1. Dp53 co-immunoprecipitates with TFIIH components. (A) Upper

panel: endogenous Dmp52 co-immunoprecipitates with Dp53 in whole cell

extracts of Dp53-FLAG S2R+ transfected cells. Immunoprecipitation (IP) was

performed using FLAG antibody. Lower panel: S2R+ protein cell extracts

were immunoprecipitated with the specific Dp53 antibody and the IP material

analyzed by western blots using the Dmp52 antibody. (B) Endogenous

Dmp52 co-immunoprecipitates with DDNp53. CoIP assays were performed

with whole cell extracts of DDNp53-V5 transfected cells and DDNp53 was

pulled down with an antibody that recognizes V5. CoIP was revealed against

other components of TFIIH with Cdk7 and XPB specific antibodies.

(C) Endogenous Dp53 co-immunoprecipitates with XPB and p62 in wild type

cell extracts. CoIP assays were performed with specific antibodies against

XPB, Cdk7 and Dp53. I, input; P1, preclearing 1; UB, unbound; B, bound.

TFIIH, p53 and JNK in apoptosis 2503

Journ

alof

Cell

Scie

nce

examine whether both Dp53 isoforms and Dmp52 interact with

each other using extracts of S2R+ cells that had been co-

transfected with both tagged proteins. We demonstrated that

FLAG-Dmp52 can interact with both isoforms of V5-tagged

Dp53 in CoIP assays (supplementary material Fig. S1). Together,

these experiments strongly suggest that Dmp53 may interact with

some TFIIH subunits, even in conditions in which DNA damage

has not been induced.

We next analysed whether Dp53 and Dmp52 can interact

directly. To achieve this goal, we expressed recombinant proteins

containing different regions of Dmp52 fused to GST in E. coli

(Fig. 2A). These GST-Dmp52 proteins were used in pulldown

experiments against the Dp53 and DDNp53 isoforms and against

constructs that cover different regions of Dmp53 expressed in

vitro in a coupled transcription translation system (Fig. 2A).

Fig. 2B shows that DDNp53 interacts with the complete Dmp52

protein. We next tested the interaction of the two Dp53 isoforms

with the complete Dmp52 protein and the C-terminal domain

(CTD-Dmp52) of Dmp52, which has been shown to be important

for the interaction with the p8 subunit of TFIIH. We also tested a

construct with a deletion of the Dmp52 CTD (Dmp52DCTD)

(Fig. 2A). We found that both polypeptides (Dp53 and DDNp53)

can interact with the complete Dmp52 protein (Fig. 2C).

Intriguingly, we also observed that Dp53 contacts the CTD

region of Dmp52 as well as the rest of the protein. Multiple

contacts in different regions between p52 and XPB have been

demonstrated (Jawhari et al., 2002), and, together with our data,

these results suggest that Dmp52 can interact with different

proteins simultaneously through different regions. We next

analysed the two isoforms containing a deletion of the CTD of

Dp53 and analysed the CTD alone for its interaction with

Dmp52. We again found that both Dp53 isoforms interact with

Dmp52 but that Dp53 CTD, which contains the complete

oligomerisation domain, does not recognise Dmp52 (Fig. 2D,

lower panel). In summary, these results demonstrate that Dp53

most likely interacts with Dmp52 through its DBD even in the

absence of genotoxic stress and that Dmp52 can be added to the

list of factors that may have direct physical contact with Dmp53.

Depletion of TFIIH in the wing imaginal disc affects growth

and cell number in the adult wing

In Drosophila, mutant alleles of the Dmp52 gene are lethal, but

heteroallelic combinations of these mutants may generate adult

organisms that are smaller than wild-type flies with melanotic

Fig. 2. Dmp52 interacts physically with both isoforms of Dp53. (A) Schematic representation of Dmp52 and Dp53 proteins. Upper panel: Dmp52 scheme

showing full length protein and fragments of recombinant GST-tagged Dmp52 used to perform pulldown assays. Dmp52 fragments include a CTD domain and

protein excluding this domain fused to GST (GST-Dmp52DCTD). Lower panel: Dp53 scheme showing full length protein of both isoforms in addition to protein

fragments transcribed and translated in vitro and marked with radioactive 35S. Fragments of Dp53 consist in CTD-domain of Dp53, which is the same in both

isoforms, and their respective fragments with truncated CTD eliminating the oligomerisation domain. (B) Pulldown assay demonstrating that recombinant

GST-Dmp52 protein interacts physically with DDNp53 isoform. Although a weak signal is detected for GST-only, DDNp53 signal is stronger on incubation with

GST-Dmp52, indicating an interaction between these proteins. (C) Pulldown assay of Dmp52 and its fragments with both isoforms of Dp53. Both Dp53 isoforms

interacts physically with full length Dmp52, the CTD domain and GST-Dmp52DCTD, indicating at least two binding sites on Dmp52. (D) Pulldown assay of

Dp53 fragments and full length Dmp52. Dmp52 interacts with Dp53 through a region that excludes its oligomerisation domain.

Journal of Cell Science 126 (11)2504

Journ

alof

Cell

Scie

nce

tumours and chromosomal instabilities (Fregoso et al., 2007).

These phenotypes are pleiotropic and may be caused by the

accumulation of defects during development; it is therefore

difficult to determine the contribution of the genetic interaction

of factors such as Dp53. Therefore, we decided to analyse the

effect of the depletion of Dmp52 in a particular tissue at a

specific time. To achieve this objective, we decided to use

dsRNA against Dmp52 specifically directed to the wing disc

using the UAS-GAL4 system. The wing disc is a monolayer

epithelium that grows several thousand-fold in mass and cell

number during larval development and is an excellent model for

the study of cell growth, proliferation and differentiation (Martın

et al., 2009). Thus, to direct the expression of Dmp52i, we first

used the MS1096 driver, which directs the expression of Gal4 in

the dorsal domain of the wing pouch; we then crossed these flies

with flies carrying dsRNA against Dmp52 (Dmp52i) from the

Vienna Resource Center. Consistent with results from our

previous work using Dmp52 mutant alleles (Fregoso et al.,

2007), we observed a reduction in the size of adult wings in flies

expressing Dmp52i compared with flies carrying only the driver

(Fig. 3). We also observed deformation of the wing shape and the

presence of extra veins in the wing (Fig. 3).

To confirm these results and to determine if the expression of

Dmp52i in a different wing disc domain generates similar

phenotypes, we used an engrailed driver (en-Gal4) that directs

the expression of Dmp52i to the posterior domain of the wing

disc (Fig. 3). A reduction in wing size and the presence of extra

veins (indicated by arrows in Fig. 3) were also observed with the

en-Gal4 driver. In addition, the posterior cross vein was so highly

reduced that it almost disappeared. Interestingly, in the case of

the en-GAL4 driver, we observed that the proportion of the wing

territories was relatively well maintained, even in regions where

Dmp52i was not expressed, although the reduction in size was

stronger in the posterior domain than in the anterior domain

(supplementary material Fig. S2). In addition, the presence of

extra veins was also detected in territories outside the driver

expression. It has been recently reported that stressful conditions

that reduce the cell size in a particular territory in the wing disc

generate a non-autonomous effect to coordinate cell growth in

adjacent cell populations (Mesquita et al., 2010; Wells and

Johnston, 2012); it appears that this scenario also applied in the

case of the depletion of the Dmp52 subunit of TFIIH in the

posterior wing disc.

To confirm that the observed phenotypes are caused by the

depletion of Dmp52 in the dorsal wing disc domain, we created

flies that contain the MS1096-Gal4 driver, Dmp52i and Gal80ts,

which is an inhibitor of Gal4. We found that the phenotype of

reduced wing size and extra wing veins reverted, confirming that

those phenotypes are caused by the expression of Dmp52i in the

wing disc (supplementary material Fig. S3A). Furthermore, we

also showed that the levels of Dmp52 protein were reduced in

salivary glands expressing Dmp52i, reinforcing the specificity

of the dsRNA (supplementary material Fig. S3B). However,

different phenotypes were observed in the wing when we

previously used the MS1096-Gal4 driver to express dsRNA

against different targets (unpublished results), indicating that the

wing phenotypes observed by depleting Dmp52 are specific.

Recently, reports have suggested that subunits of TFIIH could

associate with other complexes, independent of TFIIH. This

description is true for XPD, which has been shown to interact

with the MMS19 complex involved in chromosome segregation

(Ito et al., 2010). Thus, we evaluated whether the phenotypes

generated by the depletion of Dmp52 were due to an independent

TFIIH functional effect on Dmp52 or to a generalised effect on

TFIIH. We also decided to evaluate whether these phenotypes are

also caused by the depletion of other TFIIH subunits. We directed

the expression of a dsRNA against the Drosophila p34 subunit of

TFIIH (Dmp34 in the fly) using the MS1096 driver. The depletion

Fig. 3. Expression of dsRNA against

Dmp52 (Dmp52i) reduces wing size

and causes rising of extra veins.

Dmp52 depletion in the wing on the

dorsal domain causes a ‘bent out’

phenotype in wings (upper left-hand

panel) and a marked wing size

reduction compared with controls

(upper right-hand panel). Analysis by

light microscopy of wings Dmp52

depleted shows extra veins on the

wing blade (indicated by an arrow).

Expression of Dmp52i with en-Gal4

driver shows similar phenotypes to

those observed with MS1096 driver.

Extra veins are also formed in regions

outside of Gal4 expression (arrows).

The bars in the wing images indicate

an equivalent real size.

TFIIH, p53 and JNK in apoptosis 2505

Journ

alof

Cell

Scie

nce

of Dmp34 shows similar phenotypes to those observed for the

depletion of Dmp52, although phenotypes were more severe in

the wing when using MS1096-Gal4 driver and identical when

using the en-GAL4 driver, indicating that the observed defects in

the wing are caused by a reduction in the activity of TFIIH

(supplementary material Fig. S4). The differences in the severity

of phenotypes associated with Dmp34i and Dmp52i using

MS1096 as the driver could be due to differences in RNAi

efficiency, as was observed when the temperature was modulated

(supplementary material Fig. S4A). Furthermore, using an

eyeless-Gal4 driver that directs Gal4 expression to the eyes, we

found that similar growth defects were also observed in both

Dmp52i and Dmp34i (supplementary material Fig. S4C).

The reduction in the wing size and the appearance of extra

veins suggest that the depletion of Dmp52 may be affecting cell

proliferation and/or cell growth as well as cell differentiation. To

determine if the reduction in the wing size was due to fewer or

smaller wing cells, we counted the number of hairs in the wing;

these hairs are nonsensory apical projections that form from

single cells present in a specific area (Quijano et al., 2011).

Fig. 4A shows a comparison of control wings and Dmp52i wings.

The knockdown wing clearly has more hairs per area, a

disorganised hair pattern and deformations in the longitudinal

and cross veins (indicated by an arrow in Fig. 4A; Fig. 4B).

Likewise, the overall size of the Dmp52i wing is approximately

half the size of the control wing (Fig. 4C). In addition, the

Fig. 4. Dmp52 depletion generates a higher cell density per area and wings that have a lower number of total cells. (A) MS1096 .Dmp52i wings showed a

higher density of brittles per unit area. The arrow indicates a deformation in the wing vein. (B) Quantification of cells per unit area in Dmp52 depleted wings

reveals a higher cell density, indicating that cell size is smaller compared with controls. (C) Quantification of wing area. The blade wing is reduced by

approximately 50% in MS1096 .Dmp52i. (D) Quantification of number of total cells per wing indicates that although MS1096 .Dmp52i exhibits a higher cell

density, the number of total cells is lower if compared with the MS1096/+ control. Results in B–D are means6s.e.m.

Journal of Cell Science 126 (11)2506

Journ

alof

Cell

Scie

nce

number of cells per specific area is ,20% greater in the Dmp52i

wing than in controls, indicating that every cell is smaller and

that a defect is present in the growth of each individual cell

(Fig. 4B). However, even though the number of cells per unit

area is higher in the Dmp52i wing than in the control wing, the

total number of cells in the whole wing is lower in the Dmp52i

wing than in the control wing (Fig. 4D). Therefore, the depletion

of Dmp52 in the wing disc affects both the cell size and the total

number of cells, indicating that both cell growth and total cell

number are affected. In addition, the presence of extra veins

suggests alterations in the differentiation process that establishes

the vein pattern in the wing when TFIIH is not completely

functional.

Dmp52 interacts genetically with Dp53, and its

simultaneous depletion enhances a JNK-dependent

massive cell death

The reduction in the total cell number observed in the wings

depleted in Dmp52 may be caused either by a reduction in

proliferation or by an increase in cell death during wing

development. To determine which of these processes was

involved in the generation of this phenotype, wing discs were

analysed for apoptosis using TUNEL assays, the incorporation of

BrdU to visualise DNA replication and the detection of H3Pser10

to visualise the entry into mitosis. Interestingly, we observed that

the incorporation of BrdU in DNA and the distribution and

number of nuclei that were positive for H3Pser10 were not

affected in wing discs with reduced levels of Dmp52 at any

developmental stage (Fig. 5A and data not shown). However,

Dmp52-depleted wing discs showed higher rates of apoptotic cell

death than discs carrying only the driver (Fig. 5A). This result

suggests that the reduction in the cell number is at least in part

due to an increase in apoptosis during development of the wing

disc.

In a previous analysis of mutant alleles in Dmp52, we

discovered the presence of chromosomal aberrations in

neuroblast metaphasic chromosomes (Fregoso et al., 2007).

This defect may be related to the increase in apoptosis observed

in the wing discs that have reduced levels of Dmp52. Therefore,

we analysed the presence of aberrant chromosomes in these discs

with a loss of heterozygosity assay (LOH) using the mwh1

marker. Homozygous mwh1/mwh1 clones form groups of three

Fig. 5. Reduction of levels of Dmp52 induces

apoptosis in wings and generates chromosomal

instability. (A) BrdU incorporation, H3S10P

immunostaining and TUNEL labeling of wing

discs. Dmp52 depleted wing discs exhibit more

apoptotic bodies than controls; no changes in BrdU

or H3S10P were observed, indicating that cell cycle

progression is not perturbed. The arrow indicates

apoptotic bodies. Scale bars: 50 mm. (B) Reducing

levels of Dmp52 causes genomic instability in

wings. Control and GUS-Ctp53DN expressing

wings exhibit almost no mwh1 clones, but MS1096

.Dmp52i wings showed a higher number of mwh1

clones; interestingly, the highest number of mwh1

clones was present in MS1096 .Dmp52i, GUS-

Ctp53DN wings. (C) Quantification of mwh1 clones

per wing. The genotypes are indicated in the figure.

Results are means6s.e.m.

TFIIH, p53 and JNK in apoptosis 2507

Journ

alof

Cell

Scie

nce

bristles in the wing, and thus, when heterozygous mwh1/+

individuals suffer chromosomal rearrangements (loss of the wild-

type allele by deletions of segments or the complete loss of the

chromosome), they generate this particular phenotype (de

Andrade et al., 2004). Using this assay, we found a higher

frequency of mwh1 homozygous clones in wings of Dmp52-

depleted flies than in wings of flies heterozygous for the mwh1

allele carrying only the driver MS1096 (Fig. 5B,C). These results

indicate that when TFIIH is not completely functional, aberrant

chromosomes are generated; the increase in apoptosis in the wing

disc could therefore be generated by the presence of these

chromosomal aberrations.

Because Dp53 and Dmp52 interact directly with each other, we

decided to analyse the interaction between these two factors at

the genetic level by taking advantage of the specific apoptotic

phenotype observed in the wing when the levels of Dmp52 are

reduced. Genetic interactions between TFIIH and p53 have been

demonstrated in human cells derived from patients afflicted with

xeroderma pigmentosum who carry mutations in XPD and XPB

as well as in Drosophila XPB mutants (Robles et al., 1999; Wang

et al., 2003; Merino et al., 2002). These studies were performed

by overexpressing p53 to induce apoptosis in backgrounds

deficient for XPD or XPB. However, no studies have been

performed in vivo using mutants or RNAi that affect both

TFIIH and p53 functions simultaneously. To determine the

developmental effect of depleting both TFIIH and p53 in a

particular tissue at the same time, we expressed Dmp52i together

with a dominant negative mutant form of Dp53 (GUS-Ctp53DN)

that has been used as a bona fide inhibitor of Dp53 (Brodsky

et al., 2000; Shlevkov and Morata, 2012), again using the UAS-

GAL4 system and the MS1096 driver. The expression of the

dominant negative form of Dp53 does not have an effect on the

development of the wing (Brodsky et al., 2000) (Fig. 6C).

Intriguingly, Dmp52i in combination with the GUS-Ctp53DN

mutant enhanced the defects that were produced in the wing by

the depletion of Dmp52 (Fig. 6D). Similar results were observed

with an inducible dsRNA that covers the two Dp53 isoforms

(Dmp53i) together with the simultaneous expression of Dmp52i

(supplementary material Fig. S5A,C). These results suggest that

the depletion of Dmp52 generates defects that require a

functional Dp53 to reduce the effects caused by a defective

TFIIH.

Because the depletion of Dmp52 induces apoptosis, we

analysed the wing discs of MS1096 .Dmp52i,GUS-Ctp53DN

Fig. 6. Loss of function of Dmp52 and Dp53 causes a massive cell death in wing discs. (A) Control wings show a wild-type phenotype (top two panels) and

reduced number of apoptotic bodies on wing discs (bottom panel). (B) MS1096 .Dmp52i wings show a bent out phenotype and reduction in size as previously

described in Fig. 4 and several apoptotic bodies on wing discs. (C) MS1096 .GUS-Ctp53DN wings show a wild-type phenotype (top two panels) and reduced

number of apoptotic bodies in wing discs. (D) MS1096 .Dmp52i,GUS-Ctp53DN wings showed enhanced morphological defects compared with MS1096

.Dmp52i (top two panels) and causes a massive apoptotic cell death in wing discs. The bars in the wing images (second row) indicate an equivalent real size.

Scale bar: 50 mm (bottom panels).

Journal of Cell Science 126 (11)2508

Journ

alof

Cell

Scie

nce

flies using TUNEL assays and acridine orange staining.

Intriguingly, we observed a dramatic increase in cell death in

these discs, explaining the enhancement of the wing defects in

the Dmp52i adult flies when Dp53 was not functional (Fig. 6;

supplementary material Fig. S5B). Furthermore, the same

enhancement in apoptosis was observed in flies expressing the

Dmp34 RNAi in conjunction with the expression of GUS-

Ctp53DN (supplementary material Fig. S5C). Thus, the

inactivation of Dmp53 does not suppress cell death but instead

enhances it when Dmp52 or Dmp34 (TFIIH) is depleted. We also

evaluated DNA synthesis and the mitotic status in these flies and

did not find any difference from the control organisms

(supplementary material Fig. S6). Interestingly, when we

analysed the wings of MS1095 .Dmp52i,GUS-Ctp53DN flies

for LOH, we found a higher number of mwh1 clones than in the

wings of Dmp52i flies (Fig. 5C). These data suggest that in the

absence of Dmp52 and Dp53, there is an increase in chromosome

fragility that cannot be repaired, resulting in the generation of a

higher rate of chromosomal aberrations. This hypothesis is

supported by the fact that UV-induced DNA damage enhances

the apoptotic phenotype in flies with depleted Dp53 and Dmp52

(supplementary material Fig. S7A).

The next focus was to corroborate the hypothesis that the

enhancement of the Dmp52i phenotype by the absence of a

functional Dp53 is linked to apoptosis. We therefore first

expressed the p35 inhibitor of caspases that inhibit apoptosis,

together with Dmp52i in the wing and found that the defects in

the adult wing were increased, but the appearance of the wing

was different than that of the wings of flies in which Dmp52 and

Dp53 were simultaneously depleted (Fig. 7A). The adult wings

exhibited a large blister in which the dorsal and ventral wing

surfaces were separated. Furthermore, the expression of the

caspase inhibitor p35 abolished the apoptosis induced by the

depletion of Dmp52, indicating that the cell death generated by

the reduction of Dmp52 detected by TUNEL is caspase-

dependent and confirming that the TUNEL assay was detecting

apoptosis (Fig. 7B). We next investigated whether the

enhancement of apoptosis in MS1096 .Dmp52i;GUS-Ctp53DN

discs is also caspase-dependent by co-expressing p35. Fig. 7C

shows that blocking the caspase Drice by p35 suppressed the

massive apoptosis generated by the co-depletion of Dmp52 and

Dp53 in the wing disc. We also determined if the depletion of

Dmp52 in combination with the expression of p35 in the wing

disc caused alterations in the control of the cell cycle. Again, we

Fig. 7. Coexpression of p35 abolishes cell death, indicating a caspase-dependent pathway. (A) Coexpression of p35 in MS1096 .Dmp52i background causes

blisters and altered wing morphology. (B) TUNEL assay demonstrating that apoptosis generated in MS1096 .Dmp52i is abolished when p35 is co-expressed on

this genetic background. (C) The massive cell death detected by TUNEL on MS1096 .Dmp52i, GUS-Ctp53DN wing discs is also caspase-dependent. Apoptotic

bodies are abolished when p35 is co-expressed on MS1096 .Dmp52i, GUS-Ctp53DN wing discs. The genotypes are indicated in each panel. Scale bars: 50 mm.

TFIIH, p53 and JNK in apoptosis 2509

Journ

alof

Cell

Scie

nce

analysed the incorporation of BrdU in the S phase and of

H3Pser10 for the entry into M phase. Intriguingly, we did not

detect any change in the number of cells in the S phase or M

phase of the cell cycle (supplementary material Fig. S6). These

results suggest that when Dmp52 is not functional in the dorsal

region of the wing disc in the absence of Dp53 or when apoptosis

is inhibited, the different checkpoints are not responding to the

cellular stress present in this condition; these data suggest that

TFIIH plays a direct role in the checkpoint response.

It is known that Drosophila responds to stress stimuli through

the Dp53 and JNK pathways, thereby activating the expression of

pro-apoptotic genes (Brodsky et al., 2000; Luo et al., 2007). In

fact, it has been recently demonstrated that Dp53 and JNK

establish a feedback loop that amplifies the initial apoptotic

stimuli (Shlevkov and Morata, 2012). However, it has also been

reported in a p53 mutant context that a p53-independent

apoptosis mechanism is activated in response to ionising

radiation and that this mechanism depends on the JNK pathway

(McNamee and Brodsky, 2009). Based on these reports, we

decided to analyse the effect of the depletion of JNK (basket in

Drosophila) on apoptosis in wing discs that were also depleted of

Dmp52 and Dp53. To achieve this goal, we used the MS1096

driver to direct the expression of a dsRNA against the JNK

transcript. Interestingly, the depletion of Dmp52 increased the

levels of phosphorylated JNK (P-JNK), although apoptosis was

not completely abolished when JNK was depleted (Fig. 8A,B).

These results suggest that the increase in cell death in Dmp52-

depleted wing discs, although linked to the activation of JNK, is

not completely suppressed by the absence of JNK and does not

produce the same response as the depletion of Dp53 when

Dmp52 is not present. However, when we analysed Dmp52-,

Dp53- and JNK-depleted flies, apoptosis was practically

abolished (Fig. 8A,B; supplementary material Fig. S7B). These

results correlate with the fact that the levels of P-JNK are

dramatically enhanced in discs depleted in Dmp52 and Dmp53

(Fig. 8B). Together, these results indicate that the enhancement

of apoptosis observed in Dmp52- and Dp53-depleted wing discs

depends on the JNK pathway.

Inhibition of the ATPase activity of XPB using triptolide

phenocopies the massive apoptotic phenotype in Dmp52-

Dp53- depleted cells

Triptolide is a diterpene triepoxide derived from Tripterygium

wilfordii, a plant used in traditional medicine in China that has

been shown to have a potent antiproliferative effect on different

types of cancers in preclinical studies. Recently, it has been

demonstrated that triptolide specifically inhibits the ATPase

activity of the XPB subunit of TFIIH, accounting for most of the

pharmacological effects of this natural product (Titov et al.,

2011). Interestingly, most of the studies using triptolide report

that apoptosis is the principal cause of death in cancer cells (Liu,

2011). Therefore, we decided to determine if the treatment of

third instar larvae wing discs with triptolide in a Dp53-deficient

context phenocopies the apoptotic phenotype generated by the

simultaneous depletion of functional Dmp52 and Dp53. To

determine the effect of triptolide on third instar larvae wing discs,

Fig. 8. Massive apoptotic cell death in MS1096 .Dmp52i, GUS-Ctp53DN discs is JNK-dependent. (A) The expression of dsRNAi against Basket (Bski)

abolishes apoptosis induced in wing discs by loss of function of Dmp52 and Dp53 (lower panels), indicating that apoptosis in this background is JNK-dependent.

However, apoptosis induced in Dmp52 depleted wing discs is partially Bski-independent, shown by the observation that it is not completely abolished when Bski is

co-expressed in MS1096 .Dmp52i background (upper panels). (B) The depletion of Dmp52 activates the JNK pathway. However, the simultaneous depletion of

Dmp52 and Dp53 induce an enhancement of JNK phosphorylation. Note that no signal is detected in the control genotypes.

Journal of Cell Science 126 (11)2510

Journ

alof

Cell

Scie

nce

we incubated wing discs in vitro with different concentrations of

the compound at different times. We found that the incubation of

same-age wild-type third instar larvae wing discs generated

apoptosis in a dose- and time-dependent manner (supplementary

material Fig. S8). From these studies, we decided to incubate the

wing discs with 100 mM triptolide for 3 hours. Under these

conditions, the rate of apoptosis in all wing discs was similar to

that observed in discs expressing the dsRNA against Dmp52 in

the dorsal domain of the wing pouch using MS1096 as a driver.

We next incubated wing discs expressing the GUS-Ctp53DN

dominant-negative form of Dp53 using the MS1096 driver with

triptolide and found that as in the case of the double depletion of

Dmp52 and Dp53 in this wing compartment, an increase in

apoptosis occurred (Fig. 9). Thus, the inhibition of the ATPase

activity of the XPB subunit of TFIIH by triptolide in cells

deficient in functional Dmp53 generates the same phenotype as

when the TFIIH subunit Dmp52 and Dp53 are simultaneously

depleted. We determined if the increase in apoptosis generated by

the combined action of triptolide and the depletion of functional

Dp53 was also JNK-dependent. We found that the depletion of

JNK in wing discs treated with triptolide in a Dp53-deficient

context reduced the number of apoptotic cells (Fig. 9). Therefore,

treatment with triptolide phenocopies all of the phenotypes

produced by depleting Dmp52, Dp53 and JNK and supports the

recent report that the target for this natural product is TFIIH

(Titov et al., 2011). Reciprocally, these results also confirm that

the phenotypes observed by the depletion of Dmp52 and Dmp34

using RNAi are indeed caused by deficiencies in TFIIH activities.

DiscussionTFIIH subunits play a role in at least three mechanisms that are

critical for the cell: transcription by RNA pol II, DNA repair by

NER and the control of the cell cycle. Therefore, disruption of

TFIIH may cause a dramatic and generalised stressful situation

for the cell. The depletion of TFIIH in the Drosophila wing disc

produces wings that contain smaller and fewer cells per wing.

The defects in growth may be produced by the global reduction

of transcription, making this defect a minute-like phenotype, as

we have previously described (Fregoso et al., 2007). However,

the reduction in the total cell number is not due to a defect in cell

proliferation or in delays in the entry to mitosis. As was

documented in this work, the most obvious response to the

absence of TFIIH is the induction of apoptosis, which may

explain the deformation of the wing shape and the reduction in

Fig. 9. Triptolide phenocopies the apoptotic induction in Dp53 depleted cells. Third instar imaginal wing discs carrying the MS1096 driver or in

combination with the Dp53 dominant negative (GUS-Ctp53DN) or the GUS-CTp53DN plus the Bsk RNAi were incubated for 3 hours in insect medium (see

the Materials and Methods and supplementary material Fig. S8), alone or with DMSO (vehicle) or 100 mM of triptolide, and tested for apoptosis by TUNEL

assays. An increase in apoptotic bodies in the MS1096 .GUS-CTP53DN disc, in the GAL4 expression domain (indicated by an arrow and a red circle), is

observed. This phenotype is suppressed by the presence of the dsRNA that depletes Bsk at the same wing domain. Phenotypes are indicated in the figure.

TFIIH, p53 and JNK in apoptosis 2511

Journ

alof

Cell

Scie

nce

the total number of cells in the wings depleted of Dmp52 orDmp34.

Stimuli- and signalling-activated apoptosis is complex and

context-dependent. In Drosophila, factors that induce apoptosisunder different stress situations converge to produce an activationof caspases that cause the death of the cell (Shlevkov and Morata,

2012). Recently, it has been reported that E2F plays a dual role inapoptosis, as it is anti-death in Dp53-independent apoptosis andpro-death in Dp53-dependent apoptosis. Moreover, E2F and

Dp53 are involved in inducing apoptosis in an independentmanner during development, but both are required for apoptosiswhen DNA is damaged (Moon et al., 2008). However, in

previous works, it has been reported that the apoptosis that isgenerated by the overexpression of p53 requires the presence of afunctional TFIIH (Robles et al., 1999; Merino et al., 2002). In thisstudy, we found that when both TFIIH and Dp53 were not

functional in the wing imaginal disc, apoptosis was dramaticallyenhanced. This apoptosis is dependent on JNK, as thesimultaneous depletion of Dmp52 and Dp53 induced the

activation of JNK (P-JNK) at very high levels and thedepletion of JNK abolished apoptosis. However, it is importantto take in account that Ctp53DN or the Dp53i used in this work

affects both Dp53 isoforms, and since both are capable to induceapoptosis (Dichtel-Danjoy et al., 2013), we can not differentiateif the effects that we are observing are due to the depletion of theactivity of both Dp53 isoforms or just one.

Although Dp53 is a regulator of several processes (includingthe apoptotic response), some signals that activate apoptoticprogrammes are dependent on other proteins. The overexpression

of the tumour suppressor LKB1 or the ZBP-89 protein generatesan apoptotic cell death that is p53-independent but JNK-dependent, suggesting that this kinase may be an apoptosis

instigator (Bai et al., 2004; Lee et al., 2006). Furthermore, it hasbeen demonstrated that sirtuins can induce apoptosis mediated byJNK signalling independent of Dp53 (Griswold et al., 2008). Our

results are consistent with previous reports that demonstrate thatthe response to ionising radiation (IR) in discs deficient in Dp53and the checkpoint protein Chk1 (GRP in Drosophila) generatesa strong JNK-dependent apoptotic phenotype (McNamee and

Brodsky, 2009). In this situation, IR generates chromosomalaberrations, and the checkpoint controls mediated by Chk1 arenot functional; cells that are still able to progress in the cell

cycle but that contain chromosomal abnormalities (includinganeuploidy) are thereby generated. The activation of JNKsubsequently eliminates these cells by activating the apoptotic

programme. On the other hand, it has been recently reported thataneuploidy induces a Dp53-independent apoptosis that requiresJNK (Dekanty et al., 2012). However, the initial apoptosis thatwe observed to be caused by the depletion of Dmp52 was

partially dependent on JNK, as in the absence of JNK, apoptoticbodies could still be detected in the wing discs. In other words,our results suggest that the initial absence of Dmp52 activates an

apoptotic programme that does not completely depend on JNKbut that likely depends on Dp53. Therefore, it appears that JNK isnot a limiting factor required for the apoptosis that is induced by

the depletion of Dmp52 but that when Dp53 function iscompromised and TFIIH activities are affected, the JNKpathway is over-activated to induce apoptosis. This pathway is

likely similar to the pathway induced by IR in a Dp53- and Chk1-deficient cell. This point is supported by the demonstration thatthe absence of Dmp52 generates a loss of heterozygosity and

chromosomal aberrations (Fig. 5C) (Fregoso et al., 2007).

Somehow, this condition may be equivalent to the generation

of DNA damage by IR, and because TFIIH also participates in

DNA repair and the control of the cell cycle, it is possible that the

checkpoint controls may not be completely operational. Other

works have also demonstrated that a reduction in the number of

repair proteins such as DDB1 generates chromosomal instability

(Shimanouchi et al., 2006). Therefore, we propose that in the

cells that are deficient in TFIIH and Dp53, more chromosomal

aberrations are generated because Dp53 also participates in the

activation of DNA repair genes such as XPC and DDB2

(Adimoolam and Ford, 2002) and the response by the organism

to avoid the propagation of these defects is to over-activate the

JNK pathway and thereby induce apoptosis (Fig. 10). This model

is supported by the fact that UV irradiation enhances the

apoptosis observed in Dmp52- and Dp53-depleted cells.

Furthermore, the reported evidence that defects in DNA repair

and/or the reduction in transcription increases the levels of

activated JNK, thus inducing a sustained apoptotic phenotype,

supports this hypothesis (Roos and Kaina, 2006). Thus, the

induction of massive cell death in the Dmp52 and Dp53 double-

depleted cells may be a response to the accumulation of

chromosomal aberrations, as we demonstrated in the LOH

assay and by the reduction in transcription caused by the

depletion of TFIIH activities (Fig. 10). In addition the regulation

of the active JNK (P-JNK) is mediated by the phosphatase

Puckered (Puc); the expression of Puc is Dp53-dependent and

therefore deficient in Dp53-depleted cells (McEwen and Peifer,

2005). All of these findings support the hypothesis of an

induction of a JNK-dependent massive apoptosis in cells depleted

of TFIIH and Dp53.

In human p53-deficient cells undergoing DNA damage,

treatment with caffeine for ATR inhibition sensitises cells to

lethal Premature Chromatin Condensation, which is a process that

has been shown to be independent of p53 function (Nghiem et at.,

2001; Chanoux et al., 2009). Indeed, the elimination of apoptosis

Fig. 10. Model of TFIIH and Dmp53 interactions. (A) In a background of

Dp53-proficient cells, reducing levels of TFIIH relay a signal activating Dp53

and JNK. Activation of Dp53 could be involved in repair and to induce

apoptosis. (B) However, in a background of Dp53-deficient cells, reducing

levels of Dmp52 causes massive apoptosis on wing discs caused by an

enhancement of P-JNK. The increment in apoptotic cell death could be due to

an increment on DNA damage by the loss of function of Dmp52 and Dp53.

Journal of Cell Science 126 (11)2512

Journ

alof

Cell

Scie

nce

in wing cells depleted of Dmp52 by the expression of p35generates a different phenotype than that observed in the wings of

flies that were only Dmp52-depleted (Fig. 7). Because thedepletion of Dmp52 causes chromosomal instability, it ispossible that the absence of apoptosis allows for the proliferationof cells with chromosomal aberrations that generate this particular

phenotype. Perez-Garijo et al. (Perez-Garijo et al., 2004) havedemonstrated that the suppression of apoptosis by the caspaseinhibitor p35 in irradiated wing discs generates developmental

aberrations due to abnormal growth induced by the invasion andpersistent misexpression of wingless (wg) and decapentaplegic

(dpp) signalling. It is possible that the remaining damaged cells

that were Dmp52-depleted and that expressed the caspase inhibitorp35 continued dividing and constantly emitted Wg and Dppsignals, thus contributing to the developmental abnormalities.

As mentioned above, our results suggest that the first response

to the stress caused by the absence of a functional TFIIH is theinduction of apoptosis by Dp53 and partially by the JNKpathway, in agreement with the feedback loop between Dmp53

and JNK that has been suggested to be essential for the apoptoticresponse to stress. However, when both TFIIH and Dp53activities are depleted, this feedback loop may not be operating

because Dp53 is not present (Shlevkov and Morata, 2012); thispossibility suggests that if the feedback loop is affected by theabsence of Dp53, the JNK pathway is over activated by the DNA

repair and transcriptional stress generated by the absence ofTFIIH and Dp53.

The results presented in this study have important implicationsfor the understanding of the diseases associated with TFIIH

mutations in humans, including cancer. Furthermore, the fact thatthe wing disc cells deficient in Dmp52 and Dp53 suffer massivecell death may have implications for the treatment of cancer

cells. This observation is particularly relevant because of therecent discovery that the drug triptolide, which has potentantiproliferative and immunosuppressive activities and has been

tested to be effective against cancer in preclinical trails,specifically targets the XPB subunit of TFIIH (Titov et al.,2011). In fact, we demonstrated that triptolide treatment of wingimaginal discs that are deficient in functional Dp53 phenocopied

the JNK-dependent apoptotic phenotype observed in Dmp52-depleted discs. Most of the solid tumours in humans are deficientin p53 or in some of the modulators involved in p53-related

functions; thus, it is possible that the treatment of different typesof cancer cells with triptolide induces apoptosis through the JNKpathway, thus explaining the mechanism of the killing of cancer

cells that has been demonstrated for this natural product.Similarly, drugs such as PRIMA have been shown to inducecell death in p53-mutated cancer cell lines through a JNK

pathway (Li et al., 2005). Therefore, to search for possible targetsfor future cancer therapies, it will be important to determine ifp53-deficient cancer cells are more susceptible to apoptosis thannormal cells when TFIIH is affected.

The manner in which the phenotypes caused by the geneticinteraction between Dmp52 and Dp53 are related to the physicalinteractions between the two molecules is not clear at this point.

However, it is interesting that we can detect these physicalinteractions in fly cells without causing DNA damage, suggestinga dynamic behaviour between TFIIH and Dp53, two factors that

establish interactions with other factors involved in DNA damageand transcription. Also relevant is the report that human cellsdeficient in the TFIIH subunits XPD and XPB are deficient in the

induction of apoptosis by the overexpression of p53 (Robles et al.,1999). Our own research has shown that the same condition occurs

when Dp53 is overexpressed in flies with a mutated version of theDrosophila homologue of XPB (Merino et al., 2002). Interestingly,we found that the depletion of Dmp52 induced apoptosis, and that,in the absence of Dp53, this apoptosis is not suppressed but rather

enhanced in a JNK-dependent manner. Thus, the data concerningp53 overexpression and the lack of function of Dp53 in a TFIIH-deficient context are not easy to reconcile. The analysis of the

overexpression experiments can generate many differentinterpretations. In this particular case, because TFIIH is essentialfor the expression of most of the RNA pol II-transcribed genes and

because p53 is a transcription factor that activates the expression ofpro-apoptotic genes and physically interacts with TFIIH, thereduction in the activity of TFIIH may affect the expression of

these pro-apoptotic genes directly. An explanation may be that theexpression of genes induced by the ectopic expression of atranscription factor such as p53 is more sensitive to the generalreduction in transcription. Indeed, the incubation of fly embryos

with low doses of a-amanitin suppresses the homeotictransformations caused by the overexpression of homeotic geneswithout affecting other genes (Gutierrez et al., 2004). However, we

cannot rule out the possibility that the enhancement in apoptosis bythe simultaneous depletion of TFIIH and Dp53 may be related tothe direct interaction between these two factors. Nevertheless,

there is a clear functional link between TFIIH and p53 in humanand Drosophila cells, and these interactions are relevant not onlyfor understanding the manifestations present in humans withaffected TFIIH but also for the understanding of cancer and the

possible targeting of these interactions in future therapies.

Materials and MethodsDrosophila stocks

Flies were grown in standard meal to 25 C, unless another temperature wasindicated. The Drosophila strains include: en-gal4, MS1096bx, eyeless-Gal4, Sgs3-Gal4, UAS-DicerIII, UAS-p35, TubP-Gal80ts, GUS-Ctp53DN/TM6B, UASp53i,UAS-p53 and mwh1 were obtained from Bloomington Drosophila Stock Center.Stock generating RNAi against Dmp52 (v39069), Dmp34 (v101309) and Basket(v34138) was provided by Vienna Drosophila RNAi Center.

Pulldown assay

For interaction assays, overexpression of Dmp52 fused to GST in bacteria wasinduced with 0.4 mM IPTG during 4 h. GST-Dmp52 was purified usingGlutathione-Sepharose (Amersham). Pulldown assays were performed asreported in Valadez-Graham et al. (Valadez-Graham et al., 2012).

Cell culture, transfection and co-immunoprecipitation

Dmp52, DDNp53 and Dp53 were cloned into EcoRI/XhoI sites of pAc5.1/V5-HisAvector (Invitrogen). 36-FLAG was cloned in pAc5.1/HisA generating a carboxy-and amino-termini 36-FLAG where Dmp52, Dp53 and DDNp53 were cloned.

Drosophila S2R+ cells were maintained in Schneider medium with 10% fetalbovine serum and 100 mg streptomycin/0.25 mg amphotericin. Cells werecotransfected with 10 mg of each construction by calcium method. Forty-eighthours after transfection, the cells were collected and lysed. Lysates were clarifiedto 13,000 rpm at 4 C. CoIPs were performed as indicated in Herrera-Cruz et al.(Herrera-Cruz et al., 2012)

Antibodies

Antibodies used were Dp53: D-200 and H3 (both antibodies recognize the twoDp53 isoforms), p62 Q-19, p89 S-19 and Cdk7 ds-17 (Santa Cruz); Dp53-H3 andAnti-BrdU G3G4 (Hybridoma Bank). Dmp52 antibody was raised using acarboxy-terminal peptide (-CDVKRYWKKYSKSGV-) by New England Peptides.

TUNEL assay in wing discs

Third instar imaginal discs wing were dissected in PBS and fixed in 4%formaldehyde during 20 min. Terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) assay was performed using the In-Situ-Cell-Death-Detection-kit (Roche). For the quantitative analysis all discs were treated in the

TFIIH, p53 and JNK in apoptosis 2513

Journ

alof

Cell

Scie

nce

same conditions. The region of interest was then selected and then the numberparticles at the highest intensity (255 arbitrary units) were quantified using theImage-J software (NIH): In each image analysed a high-density thresholdcorresponding with the labeling of the TUNEL reaction was adjusted. StandardDeviations were calculated using Microsoft Excel.

Wing disc immunostainingFor BrdU immunostaining, imaginal wing disc was incubated in 75 mg/ml BrdU incomplete Schneider medium for 30 min at room temperature. Tissue was fixedwith 4% formaldehyde, permeabilised 30 min in PBS with 0.6% Triton X-100 andDNA was denatured for 30 min with 2N HCl. BrdU was detected by mouse anti-BrdU at dilution 1:5,000. For other immunostaining, wing disc was dissected inPBS, fixed in 4% formaldehyde in PBS during 30 min and permeabilised 30 minin PBS with 0.6% Triton X-100. Rabbit anti-H3Pser10 was used at 1:1,000.Secondary antibodies Alexa Fluor 568 goat anti-mouse and goat anti-rabbit(Invitrogen) were used at 1:250 dilutions. Rabbit a-PJNK (Abcam) was used 1:100.

Phenotypic analyses of wings and bristle countingFemale wings were collected in ethanol and fixated with lactic acid:ethanol 6:5and mounted. Cell density and total number cells was calculated as described byMoon et al. considering the number of hairs per defined area of 0.01 mm2 (Moonet al., 2008). Wing area was measured using ImageJ 1.28u. For determininganterior and posterior areas with en-GAL4, we considered veinL4 as the boundarybetween both compartments. A minimum of 10 wings were analyzed per genotype.Two-sided t-tests were performed to determine significant differences.

Loss of heterozygosity assayLarvae heterozygous for mwh1 were obtained by crossing MS1096;+/+; mwh1

females with Dmp52i males. All surviving flies were collected and wings weredissected and mounted. Only cells with three or more hairs were scored as themwh2/2 phenotype (de Andrade et al., 2004).

Total extracts and western blotTotal extracts from salivary glands was performed as described previously(Palomera-Sanchez et al., 2010). Gels were transferred into nitrocellulosemembranes and were revealed for western blot using an ECLchemiluminescence detection kit (Amersham).

Triptolide treatmentThird instar larval wing imaginal discs from Ore R, MS1096, MS1096; Bski,MS1096;+/+;GUS-ctp53DN or MS1096;Bski;GUS-ctp53DN dissected in PBSwere incubated in Shields and Sang M3 insect medium (Sigma) supplemented with2% fetal bovine serum and 1% penicillin-streptomycin (Gibco) in presence orabsence of 100 mM TPL (Biovision) or DMSO at 25 C for different times periods.Discs were washed once with the same medium without drug, fixed in 4%formaldehyde in PBS for 20 minutes and washed three times for 10 minutes withPBST (PBS with 0.3% Triton X-100). Then apoptosis was determined by TUNELassay as described.

AcknowledgementsWe thank Andres Saralegui for his advice in microscopy. We alsothank Dr Martha Vazquez and Dr Viviana Valadez-Graham fordiscussions about this work.

Author contributionsC.V. and G.C. performed experimental work, C.V., G.C. and M.Z.designed the project and analyzed the results and wrote the paper.

FundingThis study was supported by the Programa de Apoyo a Proyectosde Investigacion e Innovacion Tecnologica (PAPIIT)/UniversidadNacional Autonoma de Mexico (UNAM) [grant number IN 20109-3]; the Consejo Nacional de Ciencia y Tecnologıa (CONACyT)[grant number 127440]; Ixtli/UNAM; and the Fundacion MiguelAleman.

Supplementary material available online at

http://jcs.biologists.org/lookup/suppl/doi:10.1242/jcs.122721/-/DC1

ReferencesAdimoolam, S. and Ford, J. M. (2002). p53 and DNA damage-inducible expression of

the xeroderma pigmentosum group C gene. Proc. Natl. Acad. Sci. USA 99, 12985-12990.

Akhtar, M. S., Heidemann, M., Tietjen, J. R., Zhang, D. W., Chapman, R. D., Eick, D.

and Ansari, A. Z. (2009). TFIIH kinase places bivalent marks on the carboxy-terminaldomain of RNA polymerase II. Mol. Cell 34, 387-393.

Bai, L., Yoon, S. O., King, P. D. and Merchant, J. L. (2004). ZBP-89-induced

apoptosis is p53-independent and requires JNK. Cell Death Differ. 11, 663-673.

Bergoglio, V. and Magnaldo, T. (2006). Nucleotide excision repair and related human

diseases. Genome Dyn. 1, 35-52.

Bourdon, J. C., Fernandes, K., Murray-Zmijewski, F., Liu, G., Diot, A., Xirodimas,

D. P., Saville, M. K. and Lane, D. P. (2005). p53 isoforms can regulate p53transcriptional activity. Genes Dev. 19, 2122-2137.

Brady, C. A. and Attardi, L. D. (2010). p53 at a glance. J. Cell Sci. 123, 2527-2532.

Brodsky, M. H., Nordstrom, W., Tsang, G., Kwan, E., Rubin, G. M. and Abrams,

J. M. (2000). Drosophila p53 binds a damage response element at the reaper locus.

Cell 101, 103-113.

Chanoux, R. A., Yin, B., Urtishak, K. A., Asare, A., Bassing, C. H. and Brown, E. J.

(2009). ATR and H2AX cooperate in maintaining genome stability under replicationstress. J. Biol. Chem. 284, 5994-6003.

Chymkowitch, P., Le May, N., Charneau, P., Compe, E. and Egly, J. M. (2011). Thephosphorylation of the androgen receptor by TFIIH directs the ubiquitin/proteasomeprocess. EMBO J. 30, 468-479.

Coin, F., Oksenych, V. and Egly, J. M. (2007). Distinct roles for the XPB/p52 andXPD/p44 subcomplexes of TFIIH in damaged DNA opening during nucleotideexcision repair. Mol. Cell 26, 245-256.

Coin, F., Oksenych, V., Mocquet, V., Groh, S., Blattner, C. and Egly, J. M. (2008).Nucleotide excision repair driven by the dissociation of CAK from TFIIH. Mol. Cell

31, 9-20.

de Andrade, H. H., Reguly, M. L. and Lehmann, M. (2004). Wing somatic mutation

and recombination test. Methods Mol. Biol. 247, 389-412.

Dekanty, A., Barrio, L., Muzzopappa, M., Auer, H. and Milan, M. (2012).Aneuploidy-induced delaminating cells drive tumorigenesis in Drosophila epithelia.

Proc. Natl. Acad. Sci. USA 109, 20549-20554.

Di Lello, P., Miller Jenkins, L. M., Mas, C., Langlois, C., Malitskaya, E., Fradet-

Turcotte, A., Archambault, J., Legault, P. and Omichinski, J. G. (2008). p53 andTFIIE alpha share a common binding site on the Tfb1/p62 subunit of TFIIH. Proc.

Natl. Acad. Sci. USA 105, 106-111.

Dichtel-Danjoy, M.-L., Ma, D., Dourlen, P., Chatelain, G., Napoletano, F., Robin,

M., Corbet, M., Levet, C., Hafsi, H., Hainaut, P. et al. (2013). Drosophila p53

isoforms differentially regulate apoptosis and apoptosis-induced proliferation. Cell

Death Differ. 20, 108-116.

Egly, J. M. and Coin, F. (2011). A history of TFIIH: two decades of molecular biology

on a pivotal transcription/repair factor. DNA Repair (Amst.) 10, 714-721.

Esnault, C., Ghavi-Helm, Y., Brun, S., Soutourina, J., Van Berkum, N., Boschiero,

C., Holstege, F. and Werner, M. (2008). Mediator-dependent recruitment of TFIIHmodules in preinitiation complex. Mol. Cell 31, 337-346.

Fregoso, M., Laine, J. P., Aguilar-Fuentes, J., Mocquet, V., Reynaud, E., Coin, F.,

Egly, J. M. and Zurita, M. (2007). DNA repair and transcriptional deficienciescaused by mutations in the Drosophila p52 subunit of TFIIH generate developmental

defects and chromosome fragility. Mol. Cell. Biol. 27, 3640-3650.

Fuller, M. T., Regan, C. L., Green, L. L., Robertson, B., Deuring, R. and Hays, T. S.

(1989). Interacting genes identify interacting proteins involved in microtubule

function in Drosophila. Cell Motil. Cytoskeleton 14, 128-135.

Glover-Cutter, K., Larochelle, S., Erickson, B., Zhang, C., Shokat, K., Fisher, R. P.

and Bentley, D. L. (2009). TFIIH-associated Cdk7 kinase functions in phosphorylationof C-terminal domain Ser7 residues, promoter-proximal pausing, and termination byRNA polymerase II. Mol. Cell. Biol. 29, 5455-5464.

Griswold, A. J., Chang, K. T., Runko, A. P., Knight, M. A. and Min, K. T. (2008).Sir2 mediates apoptosis through JNK-dependent pathways in Drosophila. Proc. Natl.

Acad. Sci. USA 105, 8673-8678.

Gutierrez, L., Merino, C., Vazquez, M., Reynaud, E. and Zurita, M. (2004). RNApolymerase II 140wimp mutant and mutations in the TFIIH subunit XPB

differentially affect homeotic gene expression in Drosophila. Genesis. 40, 58-66.

Herrera-Cruz, M., Cruz, G., Valadez-Graham, V., Fregoso-Lomas, M., Villicana,

C., Vazquez, M., Reynaud, E. and Zurita, M. (2012). Physical and functionalinteractions between Drosophila homologue of Swc6/p18Hamlet subunit of theSWR1/SRCAP chromatin-remodeling complex with the DNA repair/transcription

factor TFIIH. J. Biol. Chem. 287, 33567-33580.

Ito, S., Kuraoka, I., Chymkowitch, P., Compe, E., Takedachi, A., Ishigami, C., Coin,

F., Egly, J. M. and Tanaka, K. (2007). XPG stabilizes TFIIH, allowing

transactivation of nuclear receptors: implications for Cockayne syndrome in XP-G/CS patients. Mol. Cell 26, 231-243.

Ito, S., Tan, L. J., Andoh, D., Narita, T., Seki, M., Hirano, Y., Narita, K., Kuraoka,

I., Hiraoka, Y. and Tanaka, K. (2010). MMXD, a TFIIH-independent XPD-MMS19protein complex involved in chromosome segregation. Mol. Cell 39, 632-640.

Jawhari, A., Laine, J. P., Dubaele, S., Lamour, V., Poterszman, A., Coin, F., Moras,

D. and Egly, J. M. (2002). p52 Mediates XPB function within the transcription/repair

factor TFIIH. J. Biol. Chem. 277, 31761-31767.

Ko, L. J., Shieh, S. Y., Chen, X., Jayaraman, L., Tamai, K., Taya, Y., Prives, C. and

Pan, Z. Q. (1997). p53 is phosphorylated by CDK7-cyclin H in a p36MAT1-

dependent manner. Mol. Cell. Biol. 17, 7220-7229.

Lee, J. H., Koh, H., Kim, M., Park, J., Lee, S. Y., Lee, S. and Chung, J. (2006). JNK

pathway mediates apoptotic cell death induced by tumor suppressor LKB1 inDrosophila. Cell Death Differ. 13, 1110-1122.

Journal of Cell Science 126 (11)2514

Journ

alof

Cell

Scie

nce

Leveillard, T., Andera, L., Bissonnette, N., Schaeffer, L., Bracco, L., Egly, J. M. andWasylyk, B. (1996). Functional interactions between p53 and the TFIIH complex areaffected by tumour-associated mutations. EMBO J. 15, 1615-1624.

Li, Y., Mao, Y., Brandt-Rauf, P. W., Williams, A. C. and Fine, R. L. (2005).Selective induction of apoptosis in mutant p53 premalignant and malignant cancercells by PRIMA-1 through the c-Jun-NH2-kinase pathway. Mol. Cancer Ther. 4, 901-909.

Liu, Q. (2011). Triptolide and its expanding multiple pharmacological functions. Int.

Immunopharmacol. 11, 377-383.Luo, X., Puig, O., Hyun, J., Bohmann, D. and Jasper, H. (2007). Foxo and Fos

regulate the decision between cell death and survival in response to UV irradiation.EMBO J. 26, 380-390.

Martın, F. A., Perez-Garijo, A. and Morata, G. (2009). Apoptosis in Drosophila:compensatory proliferation and undead cells. Int. J. Dev. Biol. 53, 1341-1347.

McEwen, D. G. and Peifer, M. (2005). Puckered, a Drosophila MAPK phosphatase,ensures cell viability by antagonizing JNK-induced apoptosis. Development 132,3935-3946.

McNamee, L. M. and Brodsky, M. H. (2009). p53-independent apoptosis limits DNAdamage-induced aneuploidy. Genetics 182, 423-435.

Merino, C., Reynaud, E., Vazquez, M. and Zurita, M. (2002). DNA repair andtranscriptional effects of mutations in TFIIH in Drosophila development. Mol. Biol.

Cell 13, 3246-3256.Mesquita, D., Dekanty, A. and Milan, M. (2010). A dp53-dependent mechanism

involved in coordinating tissue growth in Drosophila. PLoS Biol. 8, e1000566.Moon, N. S., Di Stefano, L., Morris, E. J., Patel, R., White, K. and Dyson, N. J.

(2008). E2F and p53 induce apoptosis independently during Drosophila developmentbut intersect in the context of DNA damage. PLoS Genet. 4, e1000153.

Nghiem, P., Park, P. K., Kim, Y., Vaziri, C. and Schreiber, S. L. (2001). ATRinhibition selectively sensitizes G1 checkpoint-deficient cells to lethal prematurechromatin condensation. Proc. Natl. Acad. Sci. USA 98, 9092-9097.

Okuda, M., Tanaka, A., Satoh, M., Mizuta, S., Takazawa, M., Ohkuma, Y. and

Nishimura, Y. (2008). Structural insight into the TFIIE-TFIIH interaction: TFIIE andp53 share the binding region on TFIIH. EMBO J. 27, 1161-1171.

Palomera-Sanchez, Z., Bucio-Mendez, A., Valadez-Graham, V., Reynaud, E. and

Zurita, M. (2010). Drosophila p53 is required to increase the levels of the dKDM4Bdemethylase after UV-induced DNA damage to demethylate histone H3 lysine 9.J. Biol. Chem. 285, 31370-31379.

Perez-Garijo, A. F., Martın, F. A. and Morata, G. (2004). Caspase inhibition during

apoptosis causes abnormal signalling and developmental aberrations in Drosophila.

Development 131, 5591-5598.

Quijano, J. C., Stinchfield, M. J. and Newfeld, S. J. (2011). Wg signaling via Zw3 and

mad restricts self-renewal of sensory organ precursor cells in Drosophila. Genetics

189, 809-824.

Robles, A. I., Wang, X. W. and Harris, C. C. (1999). Drug-induced apoptosis is

delayed and reduced in XPD lymphoblastoid cell lines: possible role of TFIIH in p53-

mediated apoptotic cell death. Oncogene 18, 4681-4688.

Roos, W. P. and Kaina, B. (2006). DNA damage-induced cell death by apoptosis.

Trends Mol. Med. 12, 440-450.

Shimanouchi, K., Takata, K., Yamaguchi, M., Murakami, S., Ishikawa, G.,

Takeuchi, R., Kanai, Y., Ruike, T., Nakamura, R., Abe, Y. et al. (2006).

Drosophila damaged DNA binding protein 1 contributes to genome stability in

somatic cells. J. Biochem. 139, 51-58.

Shlevkov, E. and Morata, G. (2012). A dp53/JNK-dependant feedback amplification

loop is essential for the apoptotic response to stress in Drosophila. Cell Death Differ.

19, 451-460.

Titov, D. V., Gilman, B., He, Q. L., Bhat, S., Low, W. K., Dang, Y., Smeaton, M.,

Demain, A. L., Miller, P. S., Kugel, J. F. et al. (2011). XPB, a subunit of TFIIH, is a

target of the natural product triptolide. Nat. Chem. Biol. 7, 182-188.

Valadez-Graham, V., Yoshioka, Y., Velazquez, O., Kawamori, A., Vazquez, M.,

Neumann, A., Yamaguchi, M. and Zurita, M. (2012). XNP/dATRX interacts with

DREF in the chromatin to regulate gene expression. Nucleic Acids Res. 40, 1460-

1474.

Wang, X. W., Vermeulen, W., Coursen, J. D., Gibson, M., Lupold, S. E., Forrester,

K., Xu, G., Elmore, L., Yeh, H., Hoeijmakers, J. H. et al. (1996). The XPB and

XPD DNA helicases are components of the p53-mediated apoptosis pathway. Genes

Dev. 10, 1219-1232.

Wang, Q. E., Zhu, Q., Wani, M. A., Wani, G., Chen, J. and Wani, A. A. (2003).

Tumor suppressor p53 dependent recruitment of nucleotide excision repair factors

XPC and TFIIH to DNA damage. DNA Repair (Amst.) 2, 483-499.

Wells, B. S. and Johnston, L. A. (2012). Maintenance of imaginal disc plasticity and

regenerative potential in Drosophila by p53. Dev. Biol. 361, 263-276.

Zurita, M. and Merino, C. (2003). The transcriptional complexity of the TFIIH

complex. Trends Genet. 19, 578-584.

TFIIH, p53 and JNK in apoptosis 2515